Abstract
INTRODUCTION: Cross-sectional resting-state functional magnetic resonance imaging (rsfMRI) studies have revealed altered complexity with advanced Alzheimer's disease (AD) stages. The current study conducted longitudinal rsfMRI complexity analyses in AD. METHODS: Linear mixed-effects (LME) models were implemented to evaluate altered rates of disease progression in complexity across disease groups. RESULTS: The LME models revealed complexity of the higher frequency in the CNtoMCI group (those converted from cognitively normal [CN] to mild cognitive impairment [MCI]) decayed faster over time versus CN in the prefrontal and lateral occipital cortex; complexity of the lower frequency decayed faster in AD versus CN in various frontal and temporal regions (p < 0.05 & Benjamini-Hochberg corrected with q < 0.05). DISCUSSION: Local functional brain activities decayed in the early stage of the disease, and long-range communications were impacted in the later stage. Our study demonstrated longitudinal changes in AD-related rsfMRI complexity, indicating its potential as an imaging biomarker of AD. HIGHLIGHTS: We conducted longitudinal resting state functional magnetic resonance imaging (rsfMRI) complexity analyses using the Alzheimer's Disease Neuroimaging Initiative dataset.Higher-frequency complexity in the CNtoMCI group (those transitioning from cognitively normal [CN] to mild cognitive impairment [MCI]) was found to decay faster over time compared to CN, specifically in the prefrontal and lateral occipital cortex.Lower-frequency complexity was found to decay faster in AD versus CN in various frontal and temporal regions.This study demonstrated that longitudinal changes in rsfMRI complexity could serve as a potential imaging biomarker for Alzheimer's disease.